Incorrect Menu Type for Jupyter

JupyterLab extensions
1

I am trying to debug some issues with the extension I am working on where I can compile the extension on my working machine but if I download a fresh copy from my repo I cannot compile the extension.

It seems to be an issue with the type of the Jupyter Menu class but I am unsure why I am seeing these errors only on the downloaded version and not the version of the code I am actively working on.

The exact error I am seeing is below:

src/shortcuts.ts:132:32 - error TS2739: Type 'CommandRegistry' is missing the following properties from type 'CommandRegistry': processKeyupEvent, _startModifierTimer, _clearModifierTimer, _timerModifierID

132     const menuTop = new Menu({ commands: app.commands }); // create menu-bar menu
                                   ~~~~~~~~

I tried looking at git status on my working code and the only file not being tracked is .yarnrc.yml. I tried adding this file to the repo but got another error. The file contents are here:

>> cat .yarnrc.yml 
nodeLinker: node-modules

yarnPath: .yarn/releases/yarn-3.5.0.cjs

And the error when adding .yarnrc.yml is:

>> jlpm
Internal Error: The "yarn-path" option has been set (in /Users/jasonnagy/Container/workspace-lab/remote/unicodelab-ts/.yarnrc.yml), but the specified location doesn't exist (/Users/jasonnagy/Container/workspace-lab/remote/unicodelab-ts/.yarn/releases/yarn-3.5.0.cjs).
    at i (/Users/jasonnagy/opt/anaconda3/envs/jlab4/lib/python3.12/site-packages/jupyterlab/staging/yarn.js:446:11180)
    at async t (/Users/jasonnagy/opt/anaconda3/envs/jlab4/lib/python3.12/site-packages/jupyterlab/staging/yarn.js:446:10283)

My code is available on gitlab on the package_fix branch.

1 Like
2

Type ‘CommandRegistry’ is missing the following properties from type ‘CommandRegistry’: processKeyupEvent, _startModifierTimer, _clearModifierTimer, _timerModifierID

This indicates that you have incompatible versions of the package @lumino/widget with the package that provides the application (app) as it depends on different version of @lumino/commands than @lumino/widget.

It looks like you are set the dependency in package json to the latest version of "@lumino/widgets": "^2.5.0" but your lock file includes two versions: 2.3.2:

and 2.5.0:

And you also have two versions of @lumino/commands.

See the note in Develop Extensions — JupyterLab 4.3.0rc0 documentation

the Lumino widgets system on which JupyterLab relies for communication across the application requires all packages use the same copy of the @lumino/widgets package

You may be able to solve this by running:

jlpm dlx yarn-berry-deduplicate

This is only documented in Extension Migration Guide — JupyterLab 4.3.0rc0 documentation so we probably should move this to a more prominent location.

1 Like
3

Thank you for the explanation.

The version of yarn I have been using is jlpm 1.21.1 and it is my understanding that jlpm is a pinned version of yarn. When I ran the deduplication command you gave me, I got an error saying I need to upgrade yarn. Can I upgrade and subsequently use yarn in all my commands if I started with jlpm? Or should I try to upgrade jlpm? I’m just not sure how yarn and jlpm will interact if I try to upgrade and use yarn after using jlpm.

4

Are you developing against JupyterLab 3.x? JupyterLab 4.x comes with jlpm --version 3.5.0. If you did not intend to develop against JupyterLab 3.x then you can simply update jupyterlab package and it will update jlpm for you. If you did want to develop against JupyterLab 3.x, then you will not be able to use many of the @lumino package from 2.x line either way as JupyterLab 3 uses 1.x line.

By the way, JupyterLab 3.x is no longer receiving bug fixes and will only receive critical security fixes for two more months (until the end of the year).

1 Like
5

I am developing for Jupyterlab 4. Im not sure why I have a lower version as I made a new conda env for development earlier this year as I would have assumed it would pick up on any changes like that.

Is the jupyterlab package updated through npm/yarn/jlpm or something else?

1 Like
6

Is the jupyterlab package updated through npm/yarn/jlpm or something else?

I was referring to pip/conda jupyterlab package. You can try debugging by checking which jlpm and which -a jlpm (if using a Linux/Unix with compatible version of which).

1 Like